LightNews LightNews
@maximo2100.bsky.social
150 followers 930 following 250 posts
I'm interested in science, arts, philosophy, geopolitics and many more. Here, I post about technologies related to image making (among others things, I teach in film schools/universities, mostly in Paris).
Posts Media Videos Starter Packs
maximo2100.bsky.social @maximo2100.bsky.social · 9h
phantom-video.github.io/OmniInsert/
OmniInsert: Mask-Free Video Insertion of Any Reference via Diffusion Transformer Models
maximo2100.bsky.social @maximo2100.bsky.social · 9h
www.yongshengyu.com/OmniPaint-Pa...
OmniPaint: object removal, inpainting.
maximo2100.bsky.social @maximo2100.bsky.social · 9h
zhu-zhiyu.github.io/FLUX-IR/
Image Restoration
maximo2100.bsky.social @maximo2100.bsky.social · 1d
zhuang2002.github.io/FlashVSR/
FlashVSR: AI upscaler
1
maximo2100.bsky.social @maximo2100.bsky.social · 1d
huggingface.co/papers/2510....
Scaling Instruction-Based Video Editing with a High-Quality Synthetic Dataset
maximo2100.bsky.social @maximo2100.bsky.social · 2d
huggingface.co/lovis93/next...
You can do the next scenes, keeping the character, lighting, and environment.
maximo2100.bsky.social @maximo2100.bsky.social · 3d
wimmerth.github.io/anyup/
AnyUp: Upsampling/upsaling
maximo2100.bsky.social @maximo2100.bsky.social · 4d
pbihao.github.io/projects/Dre...
github.com/dvlab-resear...
DreamOmni2: Multimodal Instruction-based Editing and Generation.
Pretty much does what nano banana does but open source.
1
maximo2100.bsky.social @maximo2100.bsky.social · 6d
github.com/9nate-drake/...
huggingface.co/VeryAladeen/...
Sec segmentation tool for comfyUI
maximo2100.bsky.social @maximo2100.bsky.social · 8d
arxiv.org/abs/2510.01119
Instant4D: 4D Gaussian Splatting in Minutes
maximo2100.bsky.social @maximo2100.bsky.social · 9d
huggingface.co/papers/2510....
DreamOmni2: Multimodal Instruction-based Editing and Generation.
maximo2100.bsky.social @maximo2100.bsky.social · 10d
huggingface.co/lovis93/next...
Next-scene-qwen-image-lora-2509 is a LoRA adapter fine-tuned on Qwen-Image-Edit (build 2509), purpose-built to generate cinematic image sequences with natural visual progression from frame to frame.
maximo2100.bsky.social @maximo2100.bsky.social · 10d
arxiv.org/abs/2506.09980
PartPacker: Efficient Part-level 3D Object Generation (Nvidia research).
Given a single input image, this method generates high-quality 3D objects with an arbitrary number of complete and semantically meaningful parts. #3Dsegmentation
huggingface.co/nvidia/PartP... (demo)
maximo2100.bsky.social @maximo2100.bsky.social · 10d
research.nvidia.com/labs/toronto...
ChronoEdit: image editing (Nvidia research)
maximo2100.bsky.social @maximo2100.bsky.social · 10d
synbol.github.io/Lumina-DiMOO/
Lumina- DiMOO: generate and edit images
maximo2100.bsky.social @maximo2100.bsky.social · 11d
github.com/MeiGen-AI/In...
InfiniteTalk: ​​Unlimited-length talking video generation​​ that supports image-to-video and video-to-video generation
maximo2100.bsky.social @maximo2100.bsky.social · 11d
hanlab.mit.edu/projects/dc-...
DC-VideoGen: speeds up AI video generation
From Nvidia
maximo2100.bsky.social @maximo2100.bsky.social · 11d
github.com/xh9998/DiffVSR
DiffVSR Ai upscaler
maximo2100.bsky.social @maximo2100.bsky.social · 11d
github.com/kijai/ComfyU...
Wan Video Wrapper - Wan animate
maximo2100.bsky.social @maximo2100.bsky.social · 11d
showlab.github.io/Code2Video/
Code2Video: A Code-centric Paradigm for Educational Video Generation
maximo2100.bsky.social @maximo2100.bsky.social · 11d
arxiv.org/abs/2510.07316
Pixel-Perfect Depth with Semantics-Prompted Diffusion Transformers.
maximo2100.bsky.social @maximo2100.bsky.social · 11d
arxiv.org/abs/2501.14607
ReferDINO: Referring Video Object Segmentation with Visual Grounding Foundations
maximo2100.bsky.social @maximo2100.bsky.social · 17d
arxiv.org/abs/2412.10419
Preference Adaptive and Sequential Text-to-Image Generation (PASTA)
Google research
maximo2100.bsky.social @maximo2100.bsky.social · 17d
aaxwaz.github.io/Ovi
Ovi: video &audio generation
1
maximo2100.bsky.social @maximo2100.bsky.social · 17d
felixtaubner.github.io/cap4d/
Cap4D: 4D face reconstruction from stills